Without preinliner, we need to tune down the cold count cutoff to merge/trim more context to limit profile size for large components. However it doesn't make sense for cold threshold to be higher than hot threshold, so we now change to use hot threshold as merging/trimming cut off instead.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Unit Tests
Unit Tests
Event Timeline
llvm/tools/llvm-profgen/ProfileGenerator.cpp | ||
---|---|---|
431 | Does it still make sense to keep this? Would profile summary builder use this value? |
llvm/tools/llvm-profgen/ProfileGenerator.cpp | ||
---|---|---|
431 | Profile summary builder still use it, but the removed tuning was only needed for merging/trimming. Now that merging/trimming changed to use hot threshold, the tuning is no longer needed either. The tuning may not general enough for a good cold threshold. |
llvm/tools/llvm-profgen/ProfileGenerator.cpp | ||
---|---|---|
431 | Sounds good. |
Does it still make sense to keep this? Would profile summary builder use this value?